The Role of Cyclin-dependent Kinase Inhibitor p27Kip1 in Anti-HER2 Antibody-induced G1 Cell Cycle Arrest and Tumor Growth Inhibition*

Cyclin-dependent kinase (CDK) inhibitor p27Kip1 binds to the cyclin E·CDK2 complex and plays a major role in controlling cell cycle and cell growth. Our group and others have reported that anti-HER2 monoclonal antibodies exert inhibitory effects on HER2-overexpressing breast cancers through G1 cell cycle arrest associated with induction of p27Kip1 and reduction of CDK2. The role of p27Kip1 in anti-HER2 antibody-induced cell cycle arrest and growth inhibition is, however, still uncertain. Here we have provided several lines of evidence supporting a critical role for p27Kip1 in the anti-HER2 antibody-induced G1 cell cycle arrest and tumor growth inhibition. Induction of p27Kip1 and G1 growth arrest by anti-HER2 antibody, murine 4D5, or humanized trastuzumab (Herceptin®) are concentration-dependent, time-dependent, irreversible, and long-lasting. The magnitude of G1 cell cycle arrest induced by trastuzumab or 4D5 is well correlated with the level of p27Kip1 protein induced. Up-regulation of p27Kip1 and G1 growth arrest could no longer be removed with as little as 14 h of treatment with trastuzumab. Anti-HER2 antibody-induced p27Kip1 protein, G1 arrest, and growth inhibition persist at least 5 days after a single treatment. The magnitude of growth inhibition of breast cancer cells induced by anti-HER2 antibody closely parallels the level of p27Kip1 induced. Induced expression of exogenous p27Kip1 results in a p27Kip1 level-dependent G1 cell cycle arrest and growth inhibition similar to that obtained with anti-HER2 antibodies. Reducing p27Kip1 expression using p27Kip1 small interfering RNA blocks anti-HER2 antibody-induced p27Kip1 up-regulation and G1 arrest. Treatment with anti-HER2 antibody significantly increases the half-life of p27Kip1 protein. Inhibition of ubiquitin-proteasome pathway, but not inhibition of calpain and caspase activities, up-regulates p27Kip1 protein to a degree comparable with that obtained with anti-HER2 antibodies. We have further demonstrated that anti-HER2 antibody significantly decreases threonine phosphorylation of p27Kip1 protein at position 187 (Thr-187) and increases serine phosphorylation of p27Kip1 protein at position 10 (Ser-10). Expression of S10A and T187A mutant p27Kip1 protein increases the fraction of cells in G1 and reduces a further antibody-induced G1 arrest. Consequently, p27Kip1 plays an important role in the anti-HER2 antibody-induced G1 cell cycle arrest and tumor growth inhibition through post-translational regulation. Regulation of the phosphorylation of p27Kip1 protein is one of the post-translational mechanisms by which anti-HER2 antibody upregulates the protein.

HER2 (human epidermal growth factor receptor 2; also known as c-neu or ErbB-2) is a key member in the epidermal growth factor receptor family (1)(2)(3)(4). HER2, as a preferred heterodimer partner for other members in epidermal growth factor receptor family, plays a critical role in epidermal growth factor receptor family signaling that is linked to a variety of cellular responses to growth factors in both normal and abnormal conditions (1)(2)(3)(4). When HER2 is overexpressed in cells, normal signaling pathways are altered, and growth control is deregulated. HER2 is overexpressed in a number of cancers, including breast, ovarian, gastric, colon, and non-small cell lung carcinomas (4,5). A humanized monoclonal antibody trastuzumab or Herceptin has been developed from the murine anti-HER2 monoclonal antibody 4D5. Trastuzumab has been successfully used in clinics to treat patients with metastatic breast cancers that overexpress HER2 (4,5). Trastuzumab treatment can produce a response rate of 10 -15% as a single agent in heavily pre-treated patients with metastatic breast cancers that overexpress HER2 (6). Trastuzumab treatment produces a response rate of 25% as a single agent in first-line management of patients with HER2 positive metastatic breast cancer (7). Trastuzumab has further been shown to enhance significantly the effectiveness of chemotherapy in patients whose tumors overexpress HER2 (4,5). The combination of chemotherapy plus trastuzumab has a much higher rate of response than chemotherapy alone (50 versus 32%) (8). The combination of trastuzumab and chemotherapy also improve the time to disease progression (7.4 versus 4.6 months) and the median response duration (9.1 versus 6.1 months) when compared with chemotherapy alone (8). The mechanisms by which trastuzumab affects growth of cancer cells and response to chemotherapy are not well understood.
One of the intracellular growth regulators that are affected by trastuzumab is cyclin-dependent kinase (CDK) 1 inhibitor p27 Kip1 . p27 Kip1 , as one of the most important CDK inhibitors during cell cycle G 1 phase, binds to the cyclin E⅐CDK2 complex and plays a major role in controlling cell cycle (9). An increase in p27 Kip1 protein causes proliferating cells to exit from the cell cycle, whereas a decrease in p27 Kip1 protein promotes quiescent cells to resume cell proliferation. p27 Kip1 protein is primarily regulated post-transcriptionally at the level of both protein translation and protein stability, although transcriptional regulation and non-covalent sequestration may also occur (9 -11). Among the post-transcriptional mechanisms, ubiquitin-proteasome proteolysis is a major pathway for regulation of p27 Kip1 protein (12). Phosphorylation of p27 Kip1 protein on threonine 187 (Thr-187) by CDK2 prepares p27 Kip1 protein for binding to ubiquitin ligase SCF Skp2 that leads to 26 S proteasome degradation (13)(14)(15). In contrast to the phosphorylation of Thr-187, the phosphorylation of p27 Kip1 protein on serine 10 (Ser-10) by human kinase interacting stathmin stabilizes p27 Kip1 protein in G 1 (16,17). p27 Kip1 protein has been reported to interact with c-Jun co-activator Jab1 (also known as CSN5), and this interaction causes nuclear export of the p27 Kip1 protein (18) and modulates c-Jun-dependent transcription (19).
Because of the major impact of p27 Kip1 in controlling cell cycle, the role of p27 Kip1 protein in human carcinogenesis has been indicated in a number of studies. Low expression of p27 Kip1 protein is associated with excessive cell proliferation and has been linked to many types of human tumors including breast cancer (10). Low expression of p27 Kip1 protein is found to correlate reversibly with HER2 overexpression via HER2-facilitated p27 Kip1 degradation (20). Low levels of p27 Kip1 protein correlates well with higher grade neoplasms and poor survival rates (10). A striking correlation between the expression of tumor suppressor PTEN and the level of p27 Kip1 protein is observed in thyroid carcinoma (21). By down-regulating p27 Kip1 protein via proteasomal degradation, oncogenic fusion protein Bcr-Abl forces fibroblasts and hematopoietic cells to divide under inappropriate conditions (22). Furthermore, targeted disruption of the p27 Kip1 gene increases body size of mice, leads to striking enlargement of multiple organs and development of pituitary tumors (23,24).
In the pre-clinical setting, our group and others (25)(26)(27)(28)(29)(30)(31) have demonstrated that anti-HER2 monoclonal antibodies exert inhibitory effects on HER2-overexpressing breast cancer through induction of G 1 cell cycle arrest associated with induction of p27 Kip1 and reduction of CDK2. However, the role of p27 Kip1 in anti-HER2 antibody-induced G 1 cell cycle arrest and growth inhibition is still uncertain. Here we have provided several lines of evidence to support a critical role for p27 Kip1 in the anti-HER2 antibody-induced G 1 cell cycle arrest and tumor growth. We have further shown that regulation of the phosphorylation of p27 Kip1 protein is one of the post-translational mechanisms by which anti-HER2 antibody up-regulates the protein.

EXPERIMENTAL PROCEDURES
Cell Culture-Two human breast cancer cell lines, SKBr3 and BT474, were obtained from the American Type Culture Collection (ATCC, Manassas, VA). SKBr3 cells were grown in complete medium containing RPMI 1640 medium (Invitrogen) supplemented with 10% fetal bovine serum (Sigma), 2 mM L-glutamine, 100 units/ml of penicillin, and 100 g/ml streptomycin in humidified air with 5% CO 2 at 37°C. BT474 cells were grown in complete medium containing Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 10% fetal bovine serum, 2 mM L-glutamine, 100 units/ml of penicillin, and 100 g/ml streptomycin. For all experiments, cells were detached with 0.25% trypsin-0.02% EDTA. For cell culture, 2-6 ϫ 10 5 exponentially growing cells were plated into 100-mm tissue culture dishes or 3 ϫ 10 3 into 96-well plates in complete medium. After culture overnight in complete medium, cells were treated with antibodies at different concentrations as indicated in each figure legend in complete medium at 37°C for the indicated time intervals.
Reagents-Antibodies reactive with phospho-Thr-187 p27 Kip1 , phospho-Ser-10 p27 Kip1 , and Jab1 were purchased from Zymed Laboratories Inc. (South San Francisco, CA). An antibody to p27 Kip1 was purchased from BD Biosciences. An antibody to c-Myc was obtained from Upstate Biotechnology, Inc. (Lake Placid, NY). A monoclonal antibody to ␤-actin was purchased from Sigma. Anti-HER2 murine monoclonal antibody 4D5 and humanized monoclonal antibody trastuzumab (Herceptin) were kindly provided by Genentech (South San Francisco, CA). Human IgG1 (hIgG) purified from plasma of patients with myelomas was obtained from Calbiochem and dialyzed against sterile cold PBS to eliminate sodium azide. Hybridoma cells specific for MOPC21, which was obtained from the ATCC, Manassas, VA was used to produce ascites fluid, and the immunoglobin was purified as reported previously (25). A Tet-Off gene expression system was purchased from Clontech. Cycloheximide was purchased from Sigma. Wild-type p27 Kip1 and its mutants (S10A and T187A) in pcDNA3.0 vector were kindly provided by Dr. K. Nakayama (Departments of Molecular and Cellular Biology and Molecular Genetics, Medical Institute of Bioregulation, Kyushu University, Fukuoka, Japan). A membrane-bound GFP expression vector pEGFP-F was purchased from Clontech. Calpain inhibitor PD151746, a broad-caspase inhibitor II, and MG132 were obtained from Calbiochem.
Anchorage-dependent Growth-Two different methods have been used to assess the anchorage-dependent growth in this study. The first one was a 96-well microplate crystal violet mitogenic assay that was modified from previous reports (25,32). Briefly, 3 ϫ 10 3 of SKBr3 cells were plated into 96-well tissue culture plates in triplicate. The cells were treated with anti-HER2 antibody (trastuzumab or 4D5) or control reagent (hIgG for trastuzumab; MOPC21 for 4D5). After incubation for 3 days, the cells were washed with PBS, fixed in 1% glutaraldehyde in PBS, and stained with 0.5% crystal violet (Sigma) in methanol. The dye was eluted with Sorenson's buffer (0.9% sodium citrate, 0.02 N HCl, and 45% ethanol), and the eluted dye was measured by a microplater reader V max (Molecular Devices, Sunnyvalle, CA) at wavelength 540 nm. A second method, a low density long-term assay, was performed in 100-mm cell culture dishes. Low cell density (BT474 cells at 1 ϫ 10 5 ; SKBr3 cells at 1.5 ϫ 10 4 ) was used when plating the cells. After overnight incubation, the cells were treated with single dose of anti-HER2 antibody (trastuzumab or 4D5) or control reagent (hIgG for trastuzumab; MOPC21 for 4D5) in complete medium for up to 1 week. No medium was changed during the assay. Cells on day 3-7 after treatment were then harvested for enumeration of cells with a Coulter counter (Coulter Electronics LTD, Miami, FL), p27 Kip1 protein detection by Western blotting, and cell cycle analysis by flow cytometry.
Anchorage-independent Growth-To determine the anchorage-independent cell growth of SKBr3 cells, a colony-forming assay in soft agar was used as reported in our previous studies (25).
Cell Cycle Analysis-Cell cycle distribution was analyzed by flow cytometry. Cells were trypsinized, washed once with PBS, and fixed overnight in 70% ethanol. Fixed cells were centrifuged at 300 ϫ g for 10 min and washed with PBS. Cell pellets were re-suspended in PBS containing 50 g/ml of RNase A and 50 g/ml propidium iodide and incubated for 20 min at 37°C with gentle shaking. Stained cells were filtered through nylon mesh (41 M) and analyzed on a Coulter flow cytometer XL-MCL (Coulter Corporation, Miami, FL) for relative DNA content based on red fluorescence levels. Doublets and cell debris were excluded from the DNA histograms. The percentages of sub-G 1 cell population were determined based on relative DNA content. The percentages of cells in different cell cycle compartments were determined using the MULTICYCLE software program (Phoenix Flow Systems, San Diego, CA).
Preparation of Total Cell Lysate and Western Immunoblot Analysis-The procedures for preparation of total protein and Western immunoblot analysis were performed as described previously (25).
Infection with Adp27-2 ϫ 10 5 of SKBr3 cells were seeded in 100-mm cell culture dishes and incubated at 37°C overnight. Cells were then coinfected with a 1:1 ratio of an adeno-X Tet-Off regulatory virus (Clontech) and Ad-myc-p27 Kip1 virus (created as described above) at different multiplicities of infection (m.o.i.) or at an m.o.i. of 20. The cells were cultured with 10% Tet-free fetal bovine serum in presence (ϩ) of absence (Ϫ) of doxycycline (1 g/ml). When applicable, doxycycline was re-added every 48 h. After 48 -72 h, cells were harvested for enumeration with a Coulter counter, Western blot analysis, and cell cycle analysis.
Small Interfering RNA (SiRNA)-To silence p27 Kip1 gene expression, a single transfection of SiRNA duplex was performed using Oligofectamine reagent (Invitrogen) according to the manufacturer's protocol. Two double-stranded RNAs with 21-mers and d(TT) overhang were selected for the ability to silence p27 Kip1 expression. p27 Kip1 SiRNA 1 corresponds to nucleotides 217 to 238 of the human p27 Kip1 coding region (AAGTACGAGTGGCAAGAGGTG). p27 Kip1 SiRNA 2 corresponds to nucleotides 139 to 160 of the human p27 Kip1 coding region (AAGCACTGCAGAGACATGGAA). The SiRNAs were synthesized at high-peformance purity by Qiagen-Xeragon (Germantown, MD). A nonrelated control SiRNA, which targeted DNA sequence AATTCTC-CGAACGTGTCACGT with no significant match in the complete human genome, was purchased from Qiagen-Xeragon that had been purified under similar condition (catalog number 80-11310).
p27 Kip1 Mutants and GFP Co-transfection-SKBr3 cells grown in 60-mm dishes were transfected with the 3.2 g of appropriate expression plasmids (wild-type p27 Kip1 , S10A p27 Kip1 , or T187A p27 Kip1 ) plus 0.8 g of pEGFP-F vector using 7 l of LipofectAMINE-2000 (Invitrogen) according to manufacturer's instructions. Cells were trypsinized and re-seeded in two identical dishes after 24 h of transfection. One dish was treated with anti-HER2 antibody 4D5, and the other was treated with diluent. After incubation for another 24 h, cells were collected and subjected to cell cycle analysis or used to prepare protein. For cell cycle analysis, cells were first fixed in 0.25% paraformaldehyde for 1.5 h on ice and then stained with PI solution as described above. Only the GFP-positive population was gated, and its cell cycle distribution was assessed. The GFP positive rate in empty vector pcDNA3.0 was used to normalize other plasmids' different transfection efficiency.
Statistical Analysis-The two-tailed Student's t test was used to compare two different groups. Values with p Ͻ 0.05 were considered significant.

Anti-HER2 Antibody Induces a Concentration-dependent Accumulation of p27 Kip1
Protein and a Corresponding Concentration-dependent G 1 Cell Cycle Arrest-In previous studies, we have shown that the anti-HER2 antibody ID5 induced G 1 cell cycle arrest through inhibition of CDK2 and induction of p27 Kip1 (25). In this study, we have employed the clinically approved anti-HER2 antibody trastuzumab and its mouse precursor 4D5. Two breast cancer cell lines that overexpress HER2 protein, SKBr3 and BT474, were used in our experiments. SKBr3 cells were treated for 24 h with anti-HER2 antibody 4D5 at different concentrations or control mouse antibody MOPC21 at 10 g/ml. Cells were then divided into two portions: one for cell cycle analysis and the other for total protein extraction and Western immunoblot analysis. As shown in Fig.  1A, 4D5 induced a dramatic and concentration-dependent increase in the fraction of cells in G 1 phase of the cell cycle. This G 1 arrest was associated with a dramatic and concentrationdependent decrease of cells in S phase. The decrease of cells in G 2 /M phase was also concentration-dependent but less impressive. The control antibody MOPC21, as expected, did not cause G 1 arrest (Fig. 1A). Importantly, G 1 arrest induced by 4D5 correlated closely with the induction of p27 Kip1 in the same dose-dependent manner (Fig. 1B). No p27 Kip1 induction was observed in cells treated with control antibody MOPC21 (Fig.  1B), which correlated well with no G 1 arrest observed in Fig.  1A. These results were further confirmed in another HER2overexpressing breast cell line, BT474. As shown in Fig. 1, C and D, 4D5 treatment resulted in a concentration-dependent increase in the fraction of cells in G 1 and a concentration-dependent induction of p27 Kip1 protein. The control antibody MOPC21 elicited no p27 Kip1 protein and did not cause any G 1 arrest (Fig. 1, C and D). These data demonstrate that anti-HER2 antibody induces a concentration-dependent p27 Kip1 accumulation and a corresponding G 1 cell cycle arrest, suggesting Anti-HER2 Antibody Induces a Time-dependent Accumulation of p27 Kip1 Protein and a Corresponding Time-dependent G 1 Cell Cycle Arrest-We next test whether the induction of p27 Kip1 protein and G 1 arrest by anti-HER2 antibody also behaves in a similarly time-dependent manner. BT474 cells were treated with the anti-HER2 antibody trastuzumab at 10 g/ml or control hIgG for different time intervals. Cells were then divided into two portions: one for cell cycle analysis and the other for total protein extraction and Western immunoblot analysis. As shown in Fig. 2A, in BT474 cells trastuzumab induced a dramatic and time-dependent increase of cells in G 1 phase within 48 h. This G 1 arrest accompanied a dramatic and concentration-dependent decrease of S phase cells. In BT474 cells, the G 1 arrest induced by trastuzumab peaked around 48 h. The control hIgG, as expected, did not cause any G 1 arrest ( Fig. 2A). Similar to the results in Fig. 1, the trastuzumabinduced G 1 arrest correlated closely with the induction of p27 Kip1 in the same time-dependent manner (Fig. 2B). Trastuzumab-induced p27 Kip1 protein peaked at 48 h in BT474 cells (Fig. 2B), which was in agreement with the G 1 cell cycle arrest observed in Fig. 2A. No p27 Kip1 induction was observed in cells treated with control hIgG (Fig. 2B), which correlated well with the level of G 1 arrest observed in Fig. 2A. These results were further confirmed in the 4D5-treated SKBr3 cells. As shown in Fig. 2, C and D, 4D5 treatment resulted in a time-dependent increase in the G 1 fraction of cells and a time-dependent induction of p27 Kip1 protein. In SKBr3 cells, both p27 Kip1 level and G 1 arrest induced by 4D5 peaked around 24 h. The control antibody MOPC21 elicited no p27 Kip1 protein nor did it cause any G 1 arrest (Fig. 2, C and D). These data demonstrate that anti-HER2 antibody induces a time-dependent p27 Kip1 accumulation and a corresponding G 1 cell cycle arrest, further suggesting a role of p27 Kip1 protein in anti-HER2 antibodyinduced G 1 arrest.
Anti-HER2 Antibody Induces an Irreversible Increase in p27 Kip1 Protein and a Corresponding G 1 Cell Cycle Arrest-To determine whether anti-HER2 antibody-induced p27 Kip1 pro-tein and G 1 cell cycle arrest are reversible or irreversible, BT474 cells were treated for different intervals with or without 10 g/ml trastuzumab. At the different intervals, cells were washed with media that lacked anti-HER2 antibody and were then replenished with normal media that contained no antibody as illustrated in Fig. 3A. After 48 h, cells were harvested for measurement of p27 Kip1 protein by Western immunoblot and of cell cycle distribution by flow cytometry. Our preliminary data and published data (25) have revealed the following: 1) that anti-HER2 antibody does not induce significant p27 Kip1 protein and G 1 arrest within 8 h upon antibody treatment, and 2) that the earliest time interval for the anti-HER2 antibody to induce significant p27 Kip1 protein and G 1 arrest is around 16 h after treatment. Therefore, we have focused the time intervals between 8 and 16 h after antibody treatment. As shown in Fig.  3B, trastuzumab, as expected, induced a dramatic expression of p27 Kip1 protein and G 1 arrest (84.9%) after 48 h treatment compared with the untreated control. Treatment with trastuzumab for 8 h induced little p27 Kip1 protein and G 1 arrest. Treatment with trastuzumab for 10 and 12 h induced greater p27 Kip1 protein and G 1 arrest, which, however, did not differ statistically from untreated control. At 14 h trastuzumab induced significant p27 Kip1 protein and statistically significant G 1 arrest (80.4%) when compared with the untreated control (67.1%). These data indicate that statistically significant G 1 cell cycle arrest and elevation of p27 Kip1 protein level induced by trastuzumab could no longer be removed after 14 h of treatment with antibody. Data in Fig. 3B have also illustrated the close correlation of the degree of G 1 arrest with the level of p27 Kip1 protein induced by anti-HER2 antibody.
Growth Inhibition Induced by Anti-HER2 Antibody Correlates with the Induced Level of p27 Kip1 Protein-To investigate whether the level of p27 Kip1 protein induced by anti-HER2 antibody correlates the magnitude of growth inhibition, we have carried out two experiments. First, a microplate growth assay has been used to test the correlation between p27 Kip1 protein and anchorage-dependent growth as reported previously (32) (see "Experimental Procedures" for details). For p27 Kip1 protein detection, BT474 cells at similar cell density to the microplate growth assay were plated into 100-mm culture dishes and treated for 48 h with 4D5 and control antibody at different concentrations. As shown in Fig. 4, 4D5 induced a concentration-dependent and anchorage-dependent growth inhibition (Fig. 4A) and a concentration-dependent p27 Kip1 accumulation in BT474 cells (Fig. 4B). Under the same conditions, control antibody MOPC21 did not induce p27 Kip1 protein at the highest concentration (Fig. 4B) and any growth inhibition (Fig.  4A). Second, an anchorage-independent assay has been employed to test the correlation between p27 Kip1 protein and growth in soft agar as reported previously (25). Again, for p27 Kip1 protein determination, BT474 cells were plated into 100-mm culture dishes under similar conditions to the growth assay and treated for 48 h with 4D5 and control antibody at different concentrations. Anti-HER2 antibody 4D5 was found to induce concentration-dependent and anchorage-independent growth inhibition of BT474 cells (Fig. 4C), whereas 4D5 induced a concentration-dependent increase in p27 Kip1 protein (Fig. 4D). Control antibody MOPC21 did not induce p27 Kip1 protein at the highest concentration (Fig. 4D) nor did it increase growth inhibition (Fig. 4C). Taken together, these results suggest that growth inhibition induced by anti-HER2 antibody correlates closely with the induced level of p27 Kip1 protein in a concentration-dependent manner.
Anti-HER2 Antibody Induces a Long-lasting Up-regulation of p27 Kip1 Protein, G 1 Cell Cycle Arrest, and Growth Inhibition-As described above and shown in Fig. 2, anti-HER2 antibodies, trastuzumab and 4D5, induced time-dependent p27 Kip1 accumulation and a corresponding G 1 cell cycle arrest up to 72 h in BT474 cells and 48 h in SKBr3 cells. To determine how long these anti-HER2 antibody's effects would last, we have investigated the effect of anti-HER2 antibody on p27 Kip1 protein, G 1 cell cycle arrest, and cell growth over a prolonged period of time. To limit cell growth to less than 75% confluency, low cell densities of cells were plated (BT474 cells at 1 ϫ 10 5 ; SKBr3 cells at 1.5 ϫ 10 4 in a 100-mm culture dish). After incubation overnight, the cells were treated with a single dose of anti-HER2 antibody (trastuzumab or 4D5) or control reagent (hIgG for trastuzumab; MOPC21 for 4D5) up to 1 week. Cells on day 3-7 after treatment were then harvested for enumeration of cells with a Coulter counter, p27 Kip1 protein detection by Western immunoblot, and cell cycle analysis by flow cytometry. As shown in Fig. 5A, trastuzumab induced an increase in p27 Kip1 protein, which was above control level on day 3 and day 5 in BT474 cells. Beyond 5 days, the levels of p27 Kip1 protein of trastuzumab-and control hIgG-treated cells were similar (data not shown). Accordingly, treatment with trastuzumab produced an average of 84.8% of BT474 cells in cell cycle G 1 fraction on day 3 and 81.3% on day 5, whereas treatment with control hIgG produced an average of 67.4% of cells in the G 1 fraction on day 3 and 68.9% on day 5 (Fig. 5B). Beyond 5 days, the percentage of G 1 fraction from trastuzumab-and control hIgG-treated cells was not significantly different (data not shown). As the result of the p27 Kip1 up-regulation and G 1 arrest, trastuzumab produced 50% anchorage-dependent growth inhibition on day 3 and day 5 compared with a hIgG control (Fig. 5C). Significant growth inhibition by trastuzumab was seen on day 6 and day 7 (data not shown).
Similarly, 4D5 induced an increase in p27 Kip1 protein, which was above the control level on day 3 and day 5 in SKBr3 cells (Fig. 5D). Beyond 5 days, the levels of p27 Kip1 protein in 4D5and control MOPC21-treated cells were not distinguishable (data not shown). Accordingly, 4D5 treatment resulted in an average of 71.2% of SKBr3 cells in the G 1 phase of cell cycle on day 3 and 68.3% on day 5, whereas control MOPC21 generated an average of 56.2% of cells in G 1 fraction on day 3 and 58.1% on day 5 (Fig. 5E). Beyond 5 days, the percentage of cells in G 1 after treatment with 4D5 and with control MOPC21 was not significantly different (data not shown). As the result of the p27 Kip1 up-regulation and G 1 arrest, 4D5 generated about 34% anchorage-dependent growth inhibition on day 3 and 50% on day 5 compared with MOPC21 control (Fig. 5F). Compared with control, significant growth inhibition by 4D5 lasted on day 6 and day 7 (data not shown). Taken together, these data demonstrate that a single dose treatment of anti-HER2 antibody induces a long-lasting p27 Kip1 up-regulation, G 1 cell cycle arrest, and growth inhibition in HER2-overexpressing breast cancer cells. These data reiterate the correlation of G 1 arrest induced by anti-HER2 antibody with the level of p27 Kip1 protein. Above data also indicate, however, that there is no timedependent correlation of growth inhibition induced by anti-HER2 antibody with the level of p27 Kip1 protein.
Induced Expression of p27 Kip1 Produces G 1 Cell Cycle Arrest and Growth Inhibition Similar to That Observed with Anti-HER2 Antibody-To demonstrate the important role of p27 Kip1 protein in anti-HER2 antibody-induced G 1 cell cycle arrest and growth inhibition, a recombinant adenovirus vector expressing a doxycycline-controlled human p27 Kip1 (Adp27, Tet-Off form) has been created. Based on our previous experience (33), we have first tested the Adp27 in SKBr3 breast cancer cells that overexpress HER2 protein. As expected, Adp27 did not express induced p27 Kip1 protein that was tagged with c-Myc in the presence of doxycycline (Fig. 6A), suggesting there was no leakage of this inducible system at indicated m.o.i. values. In the absence of doxycycline, Adp27 was induced to express p27 Kip1 protein in an m.o.i.-dependent manner (Fig. 6A), suggesting effective control of p27 Kip1 in this adenovirus system. Accordingly, induced Adp27 produced an average of 59. an average of 40% anchorage-dependent growth inhibition compared with uninduced Adp27. In addition to SKBr3 cells, we have also tested Adp27 in another HER2-overexpressing breast cancer cell line, BT474, and obtained similar results (data not shown). These observations clearly illustrated that expression of p27 Kip1 protein resulted in p27 Kip1 level-dependent G 1 cell cycle arrest and growth inhibition. Thus, these data strongly indicate the importance of p27 Kip1 protein in anti-HER2 antibody-induced G 1 cell cycle arrest and growth inhibition. Silencing Expression of p27 Kip1 Blocks Anti-HER2 Antibodymediated Induction of p27 Kip1 Protein and G 1 Arrest-To further demonstrate the important role of p27 Kip1 protein in anti-HER2 antibody-induced G 1 cell cycle arrest, we have employed the p27 Kip1 SiRNA approach to silence the expression of p27 Kip1 . If p27 Kip1 expression is critical to anti-HER2 antibodyinduced G 1 cell cycle arrest, the effect of anti-HER2 antibody on G 1 arrest should be dramatically decreased when p27 Kip1 expression is impaired. Two p27 Kip1 SiRNAs (1 and 2) that we have selected in this study were capable of effectively silencing p27 Kip1 expression in SKBr3 cells after 48 h of transfection as shown in Fig. 7A. 4D5-induced p27 Kip1 protein was completely blocked in cells treated with p27 Kip1 SiRNA 1, whereas 4D5induced p27 Kip1 protein was not significantly affected in cells treated with control SiRNA (Fig. 7B). A similar effect on 4D5induced p27 Kip1 protein was observed with p27 Kip1 SiRNA 2 (data not shown). In agreement with the level of p27 Kip1 protein, 4D5-induced G 1 arrest was also completely prevented in cells treated with p27 Kip1 SiRNA 1, whereas 4D5-induced p27 Kip1 protein was not significantly affected in cells treated with control SiRNA (Fig. 7C). Thus, the important role of p27 Kip1 protein in anti-HER2 antibody-induced G 1 cell cycle arrest is further confirmed.
Anti-HER2 Antibody Significantly Increases the Half-life of p27 Kip1 Protein-Our preliminary data obtained from Northern blots and reverse transcription-PCR suggested that anti-HER2 antibody did not increase the amount of p27 Kip1 mRNA (data not shown). Thus, anti-HER2 antibody may up-regulate p27 Kip1 protein primarily through post-translational mechanisms. To confirm that anti-HER2 antibody-induced p27 Kip1 protein is because of post-translational events, we have measured the half-life of p27 Kip1 protein. BT474 cells were treated with trastuzumab for 48 h and then treated with an inhibitor of protein synthesis, cycloheximide, at 5 g/ml, for different time intervals as indicated in Fig. 8A. Cells were harvested for Western blot analysis to check the level of p27 Kip1 protein. As shown in Fig. 8A, the level of p27 Kip1 protein in trastuzumabtreated cells declined gradually with time but at a much slower rate than in hIgG-treated cells. After normalization of p27 Kip1 expression with the loading control (␤-actin), the half-life of p27 Kip1 protein in trastuzumab-treated cells was about 5.0 h, whereas the half-life of p27 Kip1 protein in hIgG-treated cells was about 1.8 h (Fig. 8B). The prolonged half-life of p27 Kip1 protein in anti-HER2 antibody-treated cells was also confirmed by the 35 S-labeled pulse-chase experiments (data not shown). Therefore, these results illustrate that anti-HER2 antibody significantly increases the half-life of p27 Kip1 protein and suggest the accumulation of p27 Kip1 protein induced by anti-HER2 antibody results from post-translational mechanisms.
Anti-HER2 Antibody Significantly Decreases Threonine Phosphorylation of p27 Kip1 Protein at Position 187 and Increases Serine Phosphorylation of p27 Kip1 Protein at Position 10 -The results described above support the idea that anti-HER2 antibody induces p27 Kip1 expression through post-translational regulation. Ubiquitin-proteasome proteolysis plays a major role in regulating p27 Kip1 protein (12). We have tested the importance of this mechanism in SKBr3 cells using an ubiquitin-proteasome inhibitor MG 1 32. As shown in Fig. 9A, treatment with MG 1 32 significantly increased the level of p27 Kip1 protein, whereas the inhibition of calpain or caspase activity did not alter p27 Kip1 expression. These data support the important role of ubiquitin-proteasome proteolysis in regulation of p27 Kip1 protein in SKBr3 cells. Phosphorylation of p27 Kip1 protein has been widely recognized to be one of the major post-translational mechanisms that regulate the abundance of this protein (10, 12-15). Phosphorylation of p27 Kip1 protein at Threonine 187 (Thr-187) by CDK2 results in the recognition by E3 proteasome complex for protein degradation (13)(14)(15). Phosphorylation of p27 Kip1 protein at Serine 10 (Ser-10) has been shown to dramatically increase the expression of this protein (16,17). Consequently, we have investigated the phosphorylation status at these two common sites of p27 Kip1 protein using site-and phospho-specific antibodies against phosphorylated p27 Kip1 protein. As shown in Fig. 9B, anti-HER2 antibody 4D5 considerably decreased the T187 phosphorylation of p27 Kip1 protein in SKBr3 cells. At the same time, 4D5 considerably increased the Ser-10 phosphorylation of p27 Kip1 protein (Fig. 9B). The level of Jab1, a regulator of p27 Kip1 expression (18,19), decreased only slightly. Similarly, anti-HER2 antibody trastuzumab decreased Thr-187 phosphorylation and increased Ser-10 phosphorylation of p27 Kip1 protein in another breast cancer cell line, BT474 (Fig. 9C). The effect of Ser-10 phosphorylation and Thr-187 phosphorylation on anti-HER2 antibody-induced G 1 cell cycle arrest was further investigated using p27 mutants (12) and pGFP co-transfection as described under "Experimental Procedures" and in Fig. 10A. The GFP transfection efficiency for vectors expressing wildtype p27, S10A p27 (serine 10 was converted to alanine 10), and T187A p27 (threonine 187 was converted to alanine 187) ranged from 34 to 50%. Because the ratio of individual gene expression vector to GFP vector used in the transfection was 4:1, GFP-positive cells were considered to simultaneously express the individual gene of interest. As shown in Fig. 10B, S10A p27 expression increased the fraction of cells in G 1 and rendered the cells much less sensitive to 4D5-induced G 1 arrest than was observed after transfection of a control vector or wild-type p27. These results confirm Ser-10 phosphorylation is important to anti-HER2 antibody-induced G 1 arrest. Cells expressing T187A p27 also showed less sensitivity to 4D5-induced G 1 arrest (Fig. 10B). As anti-HER2 antibody decreases Thr-187 phosphorylation, the T187A mutant p27 Kip1 protein already lacks phosphorylation and should not respond to antibody. Taken together, these results showed that regulation of the phosphorylation of p27 Kip1 protein might present one of the post-translational mechanisms by which anti-HER2 antibody up-regulates the protein. The data strongly suggest the different role of Thr-187 and Ser-10 phosphorylation in regulation of p27 Kip1 protein, in which phosphorylation of p27 Kip1 protein at Thr-187 promotes protein degradation, whereas phosphorylation of p27 Kip1 protein at Ser-10 stabilizes the protein. DISCUSSION Our data demonstrate that p27 Kip1 plays a critical role in the anti-HER2 antibody-induced G 1 cell cycle arrest and tumor growth inhibition. p27 Kip1 up-regulation and corresponding G 1 cell cycle arrest induced by anti-HER2 antibody are not only concentration-dependent but also time-dependent. The magnitude of G 1 cell cycle arrest induced by anti-HER2 antibody is correlated well with the level of p27 Kip1 protein induced. The magnitude of growth inhibition of breast cancer cells treated with anti-HER2 antibody parallels closely the level of p27 Kip1 induced. Induced expression of exogenous p27 Kip1 with an inducible system results in a similar G 1 cell cycle arrest and growth inhibition to that obtained with anti-HER2 antibody. Inhibition of p27 Kip1 expression by p27 Kip1 antisense cDNA significantly impairs, but does not completely eliminate, anti-HER2 antibody-induced p27 Kip1 protein, G 1 arrest, and growth inhibition. Our data also illustrate that the majority of G 1 arrest induced by anti-HER2 antibody can no longer be removed after 14 h of treatment. Anti-HER2 antibody-induced p27 Kip1 protein, G 1 arrest, and growth inhibition last at least 5 days after a single treatment. Our data further demonstrate that anti-HER2 antibody increases the half-life of p27 Kip1 protein. Anti-HER2 antibody is able to decrease threonine phosphorylation of p27 Kip1 protein at position 187 and increase serine phosphorylation of p27 Kip1 protein at position 10. Therefore, up-regulation of p27 Kip1 by anti-HER2 antibody occurs primarily through post-translational mechanisms. Regulation of the phosphorylation of p27 Kip1 protein is identified as one of the post-translational mechanisms by which anti-HER2 antibody up-regulates the protein.
Defining the role of p27 Kip1 in the anti-HER2 antibody-induced G 1 cell cycle arrest and tumor growth inhibition is im- Ser-10 phosphorylation of p27 Kip1 protein in SKBr3 cells. SKBr3 cells were treated for 24 h with 4D5 at different concentrations or with diluent, and total proteins were prepared as described under "Experimental Procedures." Western blot analysis was performed to detect different proteins. Filters were first blotted with an anti-phospho-Thr-187 p27 Kip1 antibody and then stripped and re-probed with different antibodies in the following order: phospho-Ser-10 p27 Kip1 , Jab1, ␤-actin, and total p27 Kip1 . These results were representative of two independent experiments. C, trastuzumab decreased Thr-187 phosphorylation and increased Ser-10 phosphorylation of p27 Kip1 protein in BT474 cells. BT474 cells were treated with or without trastuzumab (10 g/ml) for 48 h. Total protein was prepared for Western blot analysis using different antibodies in the following order: phospho-Ser-10 p27 Kip1 , phospho-Thr-187 p27 Kip1 , and total p27 Kip1 .
FIG. 10. Expression of p27 Kip1 mutants causes G 1 arrest and decreases sensitivity to anti-HER2 antibody-induced G 1 arrest. A, flow cytometric analysis of GFP-positive cells. SKBr3 cells were transiently transfected with the appropriate expression plasmids (p27 Kip1 wild-type, p27 Kip1 S10A, or p27 Kip1 T187A) plus pEGFP-F vector using LipofectAMINE-2000. Cells were collected after 48 h of transfection and fixed in 0.25% paraformaldehyde for 1.5 h and then stained with propidium iodide solution. The GFP-positive population was gated. The numbers shown as an inset in the histogram were a percentage of GFPpositive cells. B, p27 Kip1 S10A and p27 Kip1 T187A mutants caused less sensitivity to anti-HER2 antibody-induced G 1 arrest than p27 Kip1 wild-type and control vector. SKBr3 cells were treated as described in A. Cells on individual dish were trypsinized and re-seeded in two identical dishes after 24 h of transfection. One dish was treated with anti-HER2 antibody 4D5, and the other was treated with diluent for another 24 h. Cells were then subjected to cell cycle analysis. The cell cycle distribution of the GFP-positive population was determined. The GFP-positive rate in empty vector pcDNA3.0 was used to normalize other plasmids' different transfection efficiencies. Representative data were shown from two separate experiments. portant. p27 Kip1 protein appears to be a critical downstream target of signaling through HER2 molecule. Elucidation of the role p27 Kip1 in the anti-HER2 antibody-induced tumor growth inhibition also indicates that any factors that affect p27 Kip1 level might also influence anti-HER2 antibody-induced tumor growth inhibition. Modulation of p27 Kip1 protein by multiple pathways might be exploited in the treatment of cancer in animal models and ultimately in clinical trials that include trastuzumab. At a minimum, levels of p27 Kip1 protein observed after trastuzumab therapy might be an intermediate biomarker for response to the antibody alone.
Expression of p27 Kip1 protein is a short event as the protein generally has short half-life as shown in Fig. 8. However, growth inhibition of HER2-overexpressing cancer cells induced by the antibody is a long-lasting event as the effect is cumulative. Therefore, the time-dependent correlation between growth and the level of p27 Kip1 protein could not be established. Recently, Yakes et al. (31) reported that antisense oligodeoxyribonucleotide of p27 Kip1 prevented trastuzumab-induced reduction of cells in S phase and induction of cells in G 1 phase, confirming our observation shown in Fig. 7.
In this study, we provide data to indicate that anti-HER2 antibody needs at least 14 h to exhibit its anti-tumor effects on G 1 arrest and p27 Kip1 up-regulation. After 14 h, 90% of the anti-tumor effect of anti-HER2 antibody can no longer be removed. Our data also indicate that a single dose of anti-HER2 antibody produces a long-lasting p27 Kip1 up-regulation and G 1 cell cycle arrest for at least 5 days and inhibits cancer cell growth for at least a week. These results are consistent with the administration of trastuzumab to patients on a weekly basis in clinic. In a severe combined immunodeficient mice model, Tokuda et al. (34) also showed that a single dose of trastuzumab inhibited about 50% tumor growth of HER2-overexpressing 4 -1ST human gastric carcinoma.
Our data support the notion that p27 Kip1 up-regulation by anti-HER2 antibody is through a post-translational mechanism. Regulation of p27 Kip1 phosphorylation by anti-HER2 antibody is one of the mechanisms leading to stabilization and accumulation of the protein. Recent reports have shown that different phosphorylation patterns of p27 Kip1 determine its binding to cyclin D1 or to cyclin E (35). Thr-187 and Ser-10 are two major sites for p27 Kip1 phosphorylation that regulate its abundance in the cell (13)(14)(15)(16)(17). Degradation of p27 Kip1 protein depends on phosphorylation of Thr-187 by cyclin E⅐CDK2 complex, transportation form the nucleus to the cytoplasm and then ubiquitination mediated by SCF Skp2 , and finally proteolysis by the 26 S proteasome (10 -15). How phosphorylation of Ser-10 increases the stability of p27 Kip1 protein is still unknown. Two new phosphorylation sites on p27 Kip1 have been identified recently. Three independent groups (36 -38) have reported that Threonine 157 of p27 Kip1 can be phosphorylated by protein kinase B/AKT. Fujita et al. (39) found that Threonine 198 of p27 Kip1 could also be phosphorylated by protein kinase B/AKT. Threonine 157 and 198 phosphorylation have been linked to regulate the cellular localization of p27 Kip1 , in that phosphorylation of these two sites blocks nuclear import of p27 Kip1 (16, 17, 36 -39). Thus, different phosphorylation events can affect not only the levels of p27 Kip1 protein but also its subcellular localization. It will be worthwhile to investigate the impact of trastuzumab treatment on Threonine 157 and 198 phosphorylation of p27 Kip1 protein.